Bioluminescence tomography (BLT) and fluorescence molecular tomography (FMT) are two new modalities to study molecular and cellular activities in small animals. These imaging tools help diagnose diseases, monitor therapies and facilitate drug designs. Geometrical and optical modeling of mice is a critically important prerequisite for BLT/FMT. The goal of this project is to develop novel mouse modeling techniques with unique features including a scheme that decomposes the mouse anatomy into different "organ regions", a deformable mouse atlas for knowledge-based segmentation, a piecewise constant fitting procedure for determination of optical parameters of organ regions to optimally match measured data from transillumination imaging experiments, and an inherent capability of compensating for the errors introduced by the diffusion approximation, geometrical errors and other factors. All these advantages contribute to improve the accuracy, efficiency and robustness of the mouse modeling for BLT. Our primary hypothesis is that the BLT reconstruction quality will be significantly improved using the proposed modeling techniques relative to the current method in which optical parameters are taken from the literature. The specific aims are to (1) build a prototype system for transillumination imaging experiments of a living mouse; (2) construct the mouse numerical models, which are equivalent to the real mice as closely as possible in terms of photon propagation, and develop a user-friendly software package for BLT applications; and (3) evaluate and validate the proposed modeling techniques in numerical simulation and phantom experiments, and demonstrate their merits in mouse studies. In this approach, the simplified mouse geometrical models and diffusion approximation models are combined to estimate representative optical parameters from measured transillumination imaging data. The mouse models will be used for BLT image reconstruction. Upon the completion of this project, the proposed mouse modeling techniques will have been validated, and the associated gains in BLT image reconstruction will have been quantified in terms of source centroid and total energy. [unreadable] [unreadable] [unreadable]